This is a good explanation. Thank youimrae wrote: ↑Tue May 04, 2021 12:08 pm The DSD/1-bit thing isn't as mad or exotic as it sounds, because a lot of traditional PCM ADCs use similar delta-sigma encoding to an oversampled 1-bit stream - before converting that to a multi-bit signal. Choosing whether to have the ADC convert to a multi-bit stream or do that conversion later is a bit like choosing to photograph in TIFF vs RAW.
Arturia Pigments 3 is out!
-
- KVRAF
- 8802 posts since 7 Oct, 2005
- Banned
- 2288 posts since 24 Mar, 2015 from Toronto, Canada
Arashi wrote: ↑Tue May 04, 2021 9:56 am384 kHz or bust! (Believe it or not, that's actually a thing now.)
Around 20 years ago there was also a competing technology called DSD (alternative to PCM) that encodes digital audio at only 1-bit, but with a sample rate around 2.8 MHz. There was a lot of arguing about which was better (which I don't understand at all), but it sort of faded away like Betamax.
https://www.youtube.com/watch?v=vu7mfxMNQWA
https://www.youtube.com/watch?v=qE2SpThKuZQ
^^ you can hear the difference. listen to harmonics and transients. especially in sound that have a lot of low end and harmonics.
Dont' get me wrong. I love my muddy sounds. I like my music to be Tom Waites meets Rolling Stones.. so will usually use 44.1/24 or 48/24 for anything that is to be rock and roll and bluesy. But if you are wanting to make modern sounding "wub wub" sounds, you probably want to try 96/24. I think it all comes down to creative choice (and also the capacity of your DAW of course!).
What are you going after sound wise.
Spotify Soundcloud Soundclick
Gear & Setup: Windows 10, Dual Xeon, 32GB RAM, Cubase 10.5/9.5, NI Komplete Audio 6, NI Maschine, NI Jam, NI Kontakt
Gear & Setup: Windows 10, Dual Xeon, 32GB RAM, Cubase 10.5/9.5, NI Komplete Audio 6, NI Maschine, NI Jam, NI Kontakt
- KVRAF
- 2942 posts since 31 Jan, 2020
I've never used 96000. I just used 44/24. I don't do any recording, i only use VST synths, fx, etc and prerecorded samples.
I have songs bounced into audio at 44/24. When i come to mix these songs do you think is it worth using 96/24? or make no difference? Obviously i'll be using EQs and reverbs etc some of which would be oversampled.
- Banned
- 2288 posts since 24 Mar, 2015 from Toronto, Canada
I am not an expert. From what I was told. You need to be working in at least 2x what your destination goal is. So if your goal is 44.1/16 you need to be working in at least 88/24 and then export to 44.1/16. But I don't think it maps quite that easily. Different hardware and software is designed with different native operating SR in mind. Some plugins and interfaces work better at 48, some at 96 some at something different. They have pretty nutty in depth reviews on audio gear at https://www.audiosciencereview.com/forum/index.php if you really want to get super detailed about it. I just go with what sounds good to me.Spring Goose wrote: ↑Tue May 04, 2021 1:18 pmI've never used 96000. I just used 44/24. I don't do any recording, i only use VST synths, fx, etc and prerecorded samples.
1/ I have songs bounced into audio at 44/24. When i come to mix these songs do you think is it worth bouncing the master to 96/24? or make no diffference? Obviously i'll be using EQs and reverbs etc some of which would be oversampled.
2/ If i change my audio interface to 96000 then i export/bounce my song should i expect the resulting audio file/s will be 96000? I'm using Focusrite Scarlett 2i2 which supports 96000, and my DAW is Bitwig. I can't see anywhere in Bitwig to choose sample rate, only my audio card device settings. (i'm testing this now, but answer the 1st question please)
The way I do it. My projects are in 48/24. Then I export my project to 48/24 wave mixdown and import it into mastering software. Then from there I export to 44/16 and upload and distribute. I recently started working in 96/24 on projects that I want to sound more "modern" sounding. Sound design experiments and so on. But I am mostly a song focused. It's all subjective really. I think there are no rules. From the readings I have done, most of my favorite recordings were all done in 44/16 .. but then again, those guys had access to professional studios and mastering engineers that have millions of dollars worth of gear.
Spotify Soundcloud Soundclick
Gear & Setup: Windows 10, Dual Xeon, 32GB RAM, Cubase 10.5/9.5, NI Komplete Audio 6, NI Maschine, NI Jam, NI Kontakt
Gear & Setup: Windows 10, Dual Xeon, 32GB RAM, Cubase 10.5/9.5, NI Komplete Audio 6, NI Maschine, NI Jam, NI Kontakt
- KVRAF
- 2942 posts since 31 Jan, 2020
Well i thought i'll give 96 a try, because i am wanting to make modern dance/electronic. Cheers pal.telecode wrote: ↑Tue May 04, 2021 1:45 pmSpring Goose wrote: ↑Tue May 04, 2021 1:18 pmI've never used 96000. I just used 44/24. I don't do any recording, i only use VST synths, fx, etc and prerecorded samples.
1/ I have songs bounced into audio at 44/24. When i come to mix these songs do you think is it worth bouncing the master to 96/24? or make no diffference? Obviously i'll be using EQs and reverbs etc some of which would be oversampled.
2/ If i change my audio interface to 96000 then i export/bounce my song should i expect the resulting audio file/s will be 96000? I'm using Focusrite Scarlett 2i2 which supports 96000, and my DAW is Bitwig. I can't see anywhere in Bitwig to choose sample rate, only my audio card device settings. (i'm testing this now, but answer the 1st question please)
I am not an expert. From what I was told. You need to be working in at least 2x what your destination goal is. So if your goal is 44.1/16 you need to be working in at least 88/24 and then export to 44.1/16. But I don't think it maps quite that easily. Different hardware and software is designed with different native operating SR in mind. Some plugins and interfaces work better at 48, some at 96 some at something different. They have pretty nutty in depth reviews on audio gear at https://www.audiosciencereview.com/forum/index.php if you really want to get super detailed about it. I just go with what sounds good to me.
The way I do it. My projects are in 48/24. Then I export my project to 48/24 wave mixdown and import it into mastering software. Then from there I export to 44/16 and upload and distribute. I recently started working in 96/24 on projects that I want to sound more "modern" sounding. Sound design experiments and so on. But I am mostly a song focused. It's all subjective really. I think there are no rules. From the readings I have done, most of my favorite recordings were all done in 44/16 .. but then again, those guys had access to professional studios and mastering engineers that have millions of dollars worth of gear.
- KVRAF
- 2825 posts since 8 Jun, 2018
it was the cheap way, but it still sounds great. i believe, and a DAC can be complicated or really simple (and perform on a par with a 7000 dollar unit, 50 dollar units), opamps, etc, i repeat; i believe, there are certainly DAC's that can't produce good sound. but nowadays, and in the past...imrae wrote: ↑Tue May 04, 2021 12:08 pm The DSD/1-bit thing isn't as mad or exotic as it sounds, because a lot of traditional PCM ADCs use similar delta-sigma encoding to an oversampled 1-bit stream - before converting that to a multi-bit signal. Choosing whether to have the ADC convert to a multi-bit stream or do that conversion later is a bit like choosing to photograph in TIFF vs RAW.
it is more a matter of taste, the discussions about opamps, for instance, it is more 'coloring' than science...
i have two DAC's that almost sound the same, and one 'DAC', a pioneer PD-7700, that sounds great on some recordings... and the other two make a mess of them... only a few recordings, cd's by the way, original, bought..
but mainly taste, my neighbour prefers the sound of the PD-7700, for instance, 1-bit stream indeed. not really high-end.. o well, if you look what rotell uses... or others..
we enter the domain of fantasy....
back to Pigments 3!
Primoridal Music: sadà\exposadà - Indusrial & Expanding Your Mind Hurts: Sound Brut
-
- KVRAF
- 35262 posts since 11 Apr, 2010 from Germany
If you use some taxing soft synths, then watch that CPU meter...Spring Goose wrote: ↑Tue May 04, 2021 2:10 pmWell i thought i'll give 96 a try, because i am wanting to make modern dance/electronic. Cheers pal.telecode wrote: ↑Tue May 04, 2021 1:45 pmSpring Goose wrote: ↑Tue May 04, 2021 1:18 pmI've never used 96000. I just used 44/24. I don't do any recording, i only use VST synths, fx, etc and prerecorded samples.
1/ I have songs bounced into audio at 44/24. When i come to mix these songs do you think is it worth bouncing the master to 96/24? or make no diffference? Obviously i'll be using EQs and reverbs etc some of which would be oversampled.
2/ If i change my audio interface to 96000 then i export/bounce my song should i expect the resulting audio file/s will be 96000? I'm using Focusrite Scarlett 2i2 which supports 96000, and my DAW is Bitwig. I can't see anywhere in Bitwig to choose sample rate, only my audio card device settings. (i'm testing this now, but answer the 1st question please)
I am not an expert. From what I was told. You need to be working in at least 2x what your destination goal is. So if your goal is 44.1/16 you need to be working in at least 88/24 and then export to 44.1/16. But I don't think it maps quite that easily. Different hardware and software is designed with different native operating SR in mind. Some plugins and interfaces work better at 48, some at 96 some at something different. They have pretty nutty in depth reviews on audio gear at https://www.audiosciencereview.com/forum/index.php if you really want to get super detailed about it. I just go with what sounds good to me.
The way I do it. My projects are in 48/24. Then I export my project to 48/24 wave mixdown and import it into mastering software. Then from there I export to 44/16 and upload and distribute. I recently started working in 96/24 on projects that I want to sound more "modern" sounding. Sound design experiments and so on. But I am mostly a song focused. It's all subjective really. I think there are no rules. From the readings I have done, most of my favorite recordings were all done in 44/16 .. but then again, those guys had access to professional studios and mastering engineers that have millions of dollars worth of gear.
- KVRAF
- 18434 posts since 16 Sep, 2001 from Las Vegas,USA
I'd be interesting in seeing someone run Pigments at 96K.....
None are so hopelessly enslaved as those who falsely believe they are free. Johann Wolfgang von Goethe
- KVRAF
- 2825 posts since 8 Jun, 2018
your preset chokes, even at 2048 sample buffer = 20 ms, doable to play (for me). standalone. 96K. also in reaper. not optimized system, but can handle softube modular quite well (is only one core..).
i am not clear, 20ms is playable.. but the preset, you dropped here, chokes at 96k... strange how the cpu meter can go to 185%...
(setting it to 256 partials, works.. not much sound different, didn't really A/B it...)
(and more partials does not always means better, it is what you want, and how the engine/generator/'oscillator' internally works. depends also on the synth... with 14 partials you get already a great sound for instance. with FM, ok... but that is the modular way... soft modular way...
less partials can give and edge, that more can't give, and yes how pigments engine works, more partials, more harmonics, more richness... but not always... o well experiment with it, after reading the manual...)
Primoridal Music: sadà\exposadà - Indusrial & Expanding Your Mind Hurts: Sound Brut
- KVRAF
- 18434 posts since 16 Sep, 2001 from Las Vegas,USA
To be honest I didn't really pay much attention to the number of Partials when I did that patch I was simply getting to know Pigments. The patch ran fine on my studio computer and I didn't pay attention to the CPU meter.WasteLand wrote: ↑Tue May 04, 2021 7:24 pm (and more partials does not always means better, it is what you want, and how the engine/generator/'oscillator' internally works. depends also on the synth... with 14 partials you get already a great sound for instance. with FM, ok... but that is the modular way... soft modular way...
less partials can give and edge, that more can't give, and yes how pigments engine works, more partials, more harmonics, more richness... but not always... o well experiment with it, after reading the manual...)
It wasn't until I moved it to this system that I noticed it caused a meltdown.
So the patch it not optimized but that was the point of posting it for use as an example of a patch generating extreme CPU demand as a test patch.
But I will say there's not much point of having the 512 Partial option if it requires a Fugaku to run it....
None are so hopelessly enslaved as those who falsely believe they are free. Johann Wolfgang von Goethe
- KVRAF
- 2825 posts since 8 Jun, 2018
what cpu has you studio computer or is it a combination with an optimized asio driver?Teksonik wrote: ↑Tue May 04, 2021 7:37 pmTo be honest I didn't really pay much attention to the number of Partials when I did that patch I was simply getting to know Pigments. The patch ran fine on my studio computer and I didn't pay attention to the CPU meter.WasteLand wrote: ↑Tue May 04, 2021 7:24 pm (and more partials does not always means better, it is what you want, and how the engine/generator/'oscillator' internally works. depends also on the synth... with 14 partials you get already a great sound for instance. with FM, ok... but that is the modular way... soft modular way...
less partials can give and edge, that more can't give, and yes how pigments engine works, more partials, more harmonics, more richness... but not always... o well experiment with it, after reading the manual...)
It wasn't until I moved it to this system that I noticed it caused a meltdown.
So the patch it not optimized but that was the point of posting it for use as an example of a patch generating extreme CPU demand as a test patch.
But I will say there's not much point of having the 512 Partial option if it requires a Fugaku to run it....
(didn't check it on my studio computer, quite older than my laptop, but outperforms it. not in cores...)
and i knew it is a test patch.
still can't rid of the feeling that already Pigments 2, at some point performed less, with a MPE patch i made. don't have backup of installers so, can't check it. and i already, do so much bug-reporting and beta testing...
o well. Syne, can run 32800 partials or so... but is of course made differently..
Primoridal Music: sadà\exposadà - Indusrial & Expanding Your Mind Hurts: Sound Brut
- KVRAF
- 2942 posts since 31 Jan, 2020
But people do use 96! It must be alright with Serum. I'm saving for a new computer.
- KVRAF
- 18434 posts since 16 Sep, 2001 from Las Vegas,USA
It's a hardly state of the art i7 8700K not overclocked, 6 core 12 thread with a base speed of 3.7 GHz and a turbo speed of 4.7GHz.
I use Focusrite's ASIO drivers. I'll go between 512 to 256 samples latency depending on what I'm doing at the time. If I'm playing my keyboard then I use 256 for 6ms latency. I always run at 44K.
Anyway here are 10 patches for Pigments. Just import from the drop down menu in Pigments.
You do not have the required permissions to view the files attached to this post.
None are so hopelessly enslaved as those who falsely believe they are free. Johann Wolfgang von Goethe
- KVRian
- 817 posts since 9 Jun, 2020
Arashi wrote: ↑Tue May 04, 2021 12:57 amI found this really interesting:telecode wrote: ↑Tue May 04, 2021 12:23 am Yeah.. you would think that is the case. I too have been told my ears are shot from my days of playing live. But I can hear the difference. I have also been told some interfaces, plugins and VSTs actually work better at 96 than at 44. 96 might have been the SR they were designed to work in. using anything lower makes them less optimized. Only thing you can do is try it and see if its true in your case.
https://www.youtube.com/watch?v=-jCwIsT0X8M
Great video - thanks for sharing that. I couldn't hear any difference between any of the sample rates - I clearly haven't got the ears of a sound engineer.